Scholarlink

What is an argument? Is an argument valid? Was a text manipulated to persuade? Since 2020, Touché fosters the development of support-technologies for decision-making and opinion-forming. To this end, the lab brings together researchers that develop systems to automatically answer questions like those above. At CLEF 2026 we do so in four tasks: (1) Fallacy Detection (new task), in which participants determine whether an argument follows a valid argument pattern; (2) Causality Extraction (new task), in which participants extract pro- and concausal claims from text; (3) Generalizability of Argument Identification in Context (new task), in which participants predict whether sentences would be annotated as an argument under different guidelines; and (4) Advertisement in Retrieval-Augmented Generation (2nd edition), in which participants detect and block advertisements in generated text. This paper details these tasks and summarizes the results of Touché 2025.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Overview of Touché 2026: Argumentation Systems

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Johannes Kiesel,
  • Marc Feger,
  • Tim Hagen,
  • Sebastian Heineking,
  • Maximilian Heinrich,
  • Maik Fröbe,
  • Katarina Boland,
  • Wilhelm Pertsch,
  • Julia Romberg,
  • Ines Zelch,
  • Stefan Dietze,
  • Matthias Hagen,
  • Martin Potthast,
  • Benno Stein

摘要

What is an argument? Is an argument valid? Was a text manipulated to persuade? Since 2020, Touché fosters the development of support-technologies for decision-making and opinion-forming. To this end, the lab brings together researchers that develop systems to automatically answer questions like those above. At CLEF 2026 we do so in four tasks: (1) Fallacy Detection (new task), in which participants determine whether an argument follows a valid argument pattern; (2) Causality Extraction (new task), in which participants extract pro- and concausal claims from text; (3) Generalizability of Argument Identification in Context (new task), in which participants predict whether sentences would be annotated as an argument under different guidelines; and (4) Advertisement in Retrieval-Augmented Generation (2nd edition), in which participants detect and block advertisements in generated text. This paper details these tasks and summarizes the results of Touché 2025.